176 



NATURE 



\yune 19, 1879 



in our early experiments, ■n-e measured the resistance of a 

 tube. The tell-tale tube ■ had to be substituted for the 

 galranometer in the ordinary Wheatstone bridge, as the 

 difference of potential between c and D fluctuated greatly 



TUBE 151- 



< 



TUBE 155 



tr 



TT 



Fig. 6. 



in the course of the experiment, causing violent swings of 

 the needle. 



A z is the battery, the A terminal of which is connected 

 at A', in the bridge arrangement, with two equal fluid 

 resistance tubes, fr and fr', of 4.20,000 ohms, placed in 

 vessels containing ice, to keep them at a constant tempe- 

 rature ; an adjustable coil resistance is inserted between 

 B and D ; the tube t t', to be tested, is"placed between D 



of the current passing. After the current in a tube has 

 commenced it is generally found that it will continue to 

 glow, even when some of the balancing resistance, B D, is 

 plugged out in the coil box, showing that when once 

 started the working resistance becomes less. If, on the 

 other hand, the current has been stopped entirely, it re- 

 quires generally a greater balancing resistance in the coil 

 box between B D to start it again than it did in the first 

 instance. After standing for a short or long time it regains 

 its normal condition, but the interval required may amount 

 to several days. The following numbers were obtained: — 



Fig. 7. 



and c, the z terminal of the battery being connected to D. 

 When the resistance is greater or less than that of the 

 tube to be tested there is an illumination in the detector 

 tube between B and C ; but when a current passes in t t', 

 balanced by a proper adjustment of the coil resistance, 

 then the glow in the detector ceases. It was ultimately 

 found that the detector tube might be suppressed because, 

 as soon as the resistance in b d is a little in excess of that 

 of the tube, the latter gives evidence by its illumination 



Subsequently we found it to be more convenient not to 

 make special determinations of the resistances of the 

 tubes beforehand in the way just described, but to obtain 

 them by reproducing the deflection of a galvanometer by 

 substituting wire resistances for the tube, or by measure- 

 ments taken with an electrometer in the manner 

 to be described, while observing the phenomena 

 of stratification. 



From measurements thus made with a tube 

 having several rings about i inch apart (No. 25, 

 like in Fig. 9), or a Spottiswoode tube with a 

 shifting terminal (No. 147, Fig. 8), we found that 

 the resistance of a vacuum tube, unlike that of 

 a wire, does not increase in the ratio of the 

 distance between the terminals for the same gas 

 at the same pressure. 

 ^ In making these experiments it was noticed 



^ that the resistance for equal distances appeared 

 to be greater in proximity with the negative pole 

 than in other parts of the tube, and fresh experi- 

 ments were in consequence undertaken to ascer- 

 tain the potential at the several rings by means 

 of a delicate Thomson-Becker quadrant electro- 

 meter furnished with an induction plate, I, Fig. 

 9, which may be adjusted to any required dis- 

 tance from the quadrant beneath it.^ The tubes 

 employed among others were No. 25, described 

 above, and two other longer tubes, namely. No. 

 149 (COj) with 12 rings 2 inches apart, and 

 No. 150 (CO2) with 17 rings also 2 inches distant. The 

 current was led through a metallic resistance to the first 

 ring, the last ring and the other pole of the battery being 

 to earth. It was found that the greatest difference of 

 potential occurs between the last ring and the last but one 

 on the negative side, the next greatest difference being 

 between the last and the last but one on the positive side, 

 but the difference in the former case is far greater than 

 in the latter; in some cases there is little or no difference 



TUBE. Kl. 



'^ ^>j^yj^K)jojinkkJ^iWimk>}j^^M\}J\\\^^ 



Fig. 8. 



in the last but one and the last but two on the negative 

 side ; in these cases the last but one on the negative side 

 was dark, while all the others had a luminosity about 

 them. The difference of potential between the rest of 

 the rings is sensibly uniform. 



The following observations, made December 21, 1877, 

 with tube 150, may be taken as an illustration of the 

 method of measurement adopted. Batteries 6 and 7 



' A tube selected for the readiness with which it permits the passage of a 

 ctirrent of 440 cells. 



(2,400 rod cells) were employed, and adjustable resist- 

 ances were inserted in circuit for the double purpose of 

 affording the means of readily varying the strength of 

 current without interruption, and of enabling a measure- 

 ment of that current to be made with the electrometer. 

 The connections are shown in the diagram, Fig. 9. 



* It has since been found more advantageous to separate the induction 

 apparatus from the electrometer. Each pair of quadrants is charged with 

 opposite electricities by means of two separate batteries of twenty chloride of 

 silver cells, the opposite poles being to earth ; and the induced plate of the 

 induction apparatus communicates with the needle. 



